Vol. 2, 1039-1048, Juite 1996 Clinical Cancer Research 1039
Characterization of a Novel Androgen-sensitive, Prostate-specific
Antigen-producing Prostatic Carcinoma Xenograft: LuCaP 23’
William J. Ellis,2 Robert L. Vessella,Kent R. Buhier, Franck Bladou,
Lawrence D. True, Steven A. Bigler, Dena Curtis,
and Paul H. Lange
Department of Urology, University of Washington, Seattle,
Washington 98195 and Section of Urology, Seattle Veterans Affairs
Medical Center. Seattle, Washington 98108
ABSTRACT
Prostatic carcinoma has proven extremely difficult to
establish as cell lines or xenografts. In this article, we de-
scribe a new series of prostate cancer xenografts propagated
in athymic mice, designated LuCaP 23, developed from
prostate metastases harvested at autopsy shortly after death.
Tumor from three separate metastatic deposits was devel-oped into three xenograft sublines: two from lymph nodemetastases (LuCaP 23.1 and 23.8) and one from a liver
metastasis (LuCaP 23.12). Fluorescence in situ hybridization
analysis confirms the xenografts are human. Histologically,
the xenografts are comprised of columnar epithelial cells
arranged in a glandular pattern. Tumor doubling times
range from 1 1 to 21 days for the three sublines. The cells
secrete large amounts of prostate-specific antigen (PSA)
with PSA indices of 1.27, 1.63, and 5.2! ng/mI/mm3 for themice bearing the LuCaP 23.1, 23.8, and 23.12 sublines,
respectively. Following androgen deprivation a temporary
decrease in PSA secretion and a decrease in tumor size are
noted in most tumors. Eventually, the tumors become an-
drogen independent and resume growth in castrate hosts.
The degree of PSA response to castration and time to PSAnadir correlate with time to progression. Thus, unlike most
existing models of prostatic carcinoma, this novel xenograft
exhibits many phenotypic characteristics of clinical prostatic
carcinoma, including androgen sensitivity. These properties
make this xenograft an excellent model for future study.
INTRODUCTION
Carcinoma of the prostate is the most commonly diagnosed
malignancy and the second leading cause of cancer-related
death in American men ( 1 ). Insight into the biological behavior
of this disease lags behind that of many other solid tumors. In
part, this deficiency is due to the lack of adequate in vito and in
vitro models available for study, especially when compared to
other common tumors such as breast, lung, and colon. At this
time, only eight continuously passaged prostate cancer cell lines
(2-10) and nine serially passable xenografts lines (both primary
and metastatic) (9, 1 1-17) currently exist (summarized in
Table I).
Most of the above-mentioned xenografts and cell lines are
considered inadequate models for study because they lack the
multiplicity of characteristics that make for a comprehensive
model, i.e., androgen sensitivity, production of PSA3 and PAP,
the ability to grow at a rate that allows timely experimentation,
and growth without manipulation of hormones or growth fac-
tors. Androgen sensitivity and PSA production are two hall-
marks of the prostatic phenotype, both benign and malignant,
which are of clinical importance. Historically, hormonal manip-
ulation has played a major role in the treatment of prostatic
carcinoma. Today, PSA is widely used to both diagnose prostate
carcinoma and monitor the effect of therapy on the disease.
Thus, androgen-sensitive, PSA-producing models are desirable
because they allow scientists to model clinical scenarios.
In this article, we describe a new series of prostate cancer
xenografts derived from multiple metastases obtained at au-
topsy. These xenografts, designated the LuCaP 23 series, share
many characteristics of adenocarcinoma of the prostate, and we
believe they represent an excellent model for study of this
disease.
MATERIALS AND METHODS
Clinical History. The donor of tissue for establishment
of the LuCaP 23 series was a 63-year-old white male diagnosed
with a stage DI adenocarcinoma of the prostate with a Gleason
grade of 3 + 5. He was treated with external beam radiation
therapy to the pelvis. Within 6 months of completion of radia-
tion therapy, evidence of bony metastases was discovered. The
patient was initially managed with bicalutamide (Casodex)
monotherapy. When evidence of disease progression was noted
within a few months, bilateral orchiectomy was performed.
Shortly thereafter, his disease became hormone refractory and
liver metastases were detected. The patient was then treated with
both Adriamycin and carboplatinum. He continued to show
evidence of disease progression with extensive retroperitoneal
and liver disease. The patient expired several months later with
a serum PSA level of approximately 8000 ng/ml. Within 2 h of
death, an autopsy utilizing aseptic technique was performed.
Metastatic tumor foci from several retroperitoneal lymph nodes
and the liver were harvested and returned to our laboratory for
evaluation.
Received 9/8/95: revised 2/2 1/96: accepted 2/22/96.
� Supported in part by the Lucas Foundation and a George M. O’Brien
Prostate Cancer Research Center (P050DK476S6).
2 To whom requests for reprints should be addressed, at Department ofUrology, Box 356510, University of Washington. Seattle, WA 98195.
3 The abbreviations used are: PSA, prostate-specific antigen: PAP. pros-
tatic acid phosphatase; EGF, epidermal growth factor; DHT. di-hydro-l7�3-testosterone; BPH. benign prostatic hyperplasia: FISH. fluores-cence in situ hybridization.
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1040 LuCaP 23 Prostatic Cancer Xenograft
Table 1 Characteri sties of existing p rostate cancer xenografts
Tumorigenic AndrogenName Derived from: in mice sensitive PSA produced Year derived Reference
Established prostate cancer cell lines
EB-33 Prostate Yes No No 1974 2
DU-14S Brain Yes No No 1977 3
PC-3 Vertebrae Yes No No 1976 4LNCAP Lymph node ± Yes Yes 1980 5
PC-93 Prostate No No No 1983 6TSU-PR1 Lymph node Yes No No 1987 7
JCA- 1 Prostate Yes No No 1990 8DUPRO- 1 Lymph node Yes No No 1991 9ND-i Prostate Yes No Trace 1991 10
Established prostate cancer xenografts
PC-82 Prostate Yes Yes Yes 1977 11
Honda Testis Yes Yes No 1977 12
9479 Bone Yes No No 1981 13PC-EW Lymph node Yes Yes Yes I 98 1 14PC-l33 Bone Yes No No 1981 15PC-135 Prostate Yes No No 1982 15
PC-EG Prostate Yes Yes Yes 1984 16TEN/12 Prostate Yes Yes No 1985 17
DU5683 Lymph node Yes Yes Yes 1982 9
LuCaP 23.1 Lymph node Yes Yes Yes 1991
LuCaP 23.8 Lymph node Yes Yes Yes 1991
LuCaP 23. 12 Liver Yes Yes Yes 1991
Establishment of Primary Cell Lines in Vitro. Primary
tissue culture was artempted for each lymph node and liver
metastatic foci by placing 1-2-mm3 tissue fragments of tumor
on an etched X in a 100-mm culture dish. The fragments were
bathed in RPMI 1640 (BioWhittaker, Walkersville, MD) sup-
plemented with 15% heat-inactivated fetal bovine serum (Inter-
gen, Purchase, NY), EOF (10 ng/ml) (Collaborative Research,
Bedford, MA), and DHT (0.1 ng/ml; Sigma Chemical, St. Louis,
MO). In one half of the culture dishes, ascitic fluid from the
patient was added to the media at a 1:4 ratio. For passage, cells
were removed using 0.25% trypsin, washed with 1 X PBS, and
resuspended in the media described above.
Attempts were made to establish primary cell lines from
xenografts (vide infra) using a variety of different protocols. The
first attempt was made using the identical protocol as the pri-
mary tissues. Cellular dissociation was done both mechanically
and enzymaticaliy. Xenograft tissue was cut into small bits,
passed through a cell sieve, then digested with 0. 1% coilagenase
type IV, 0.01 % hyaluronidase, and 0.002% DNase (all from
Sigma Chemical) for up to 12 h at 4#{176}C.Cells were separated
from debris by layering over Ficoll-Hypaque 400 (Accurate
Chemical, Westbury, NY) and centrifuging at X 100. Viable
cells were plated at high densities in RPM! 1640 and DMEIF12
media (BioWhittaker) both supplemented with 10% fetal bovine
serum, EOF (10 ng/mi; Collaborative Research), DHT (1 X
l0_8 sdml; Sigma), bovine pituitary extract (30 p.g/ml; Collab-
orative Research), and insulin-transferrin-selenium (20 �i.g/20
�i�g/5 ng/ml; Sigma). Other treatments of xenograft tissue in-
elude plating on collagen (Collaborative Research), fibronectin
(Collaborative Research), Pronectin F (Protein Polymer Tech-
nologies, San Diego, CA), or thin-layer Matrigel (Collaborative
Research)-coated tissue culture vessels. Cells were also sus-
pended in Matrigel spaghetti using the technique of Gingrich et
al. (9) and floated in the media described above.
Animals. All animal use and procedures were approved
and done in accordance with the University of Washington and
Seattle Veterans Affairs Animal Care Committee Guidelines.
Xenografts were maintained in BALB/C-nu/nu (athymic) mice
(Simonsen Laboratories, Oilroy, CA). Mice were housed in
microisolator caging placed in forced HEPA-filtered air shelv-
ing units and maintained at 23.8#{176}Croom temperature, relative
humidity 27-45%, on a 10-h Iight/l4-h dark-light cycle. Picolab
Irradiated Rodent Chow 20 (PM! Feeds, Inc.) and autoclaved
water were provided ad libitum. s.c. tumor implantation and
surgical castration were performed under ketamine/xylazine an-
esthesia (260.0 mgIl7.6 mg/kg). Wounds were closed with
9-mm wound clips. Animals were sacrificed when tumor vol-
ume reached 1000 mm3.
Establishment of the Xenografts. Tissue from autopsy
that was judged to be suitable for in vivo growth, three tissue
samples from retroperitoneal lymph nodes and one from liver,
was washed in 1 X PBS and then soaked for 20 mm in 2.5 �.tg/ml
amphotericin (Life Technologies, Inc., Grand Island, NY) and
50 p.g/ml gentamicin (Life Technologies, Inc.). Tumor bits
(approximately 20-25 mm3) were then implanted s.c. in
6-week-old male athymic mice. From each metastatic foci,
tumor bits were randomized among sites and animals to remove
individual animal effects, giving a possible 12 sublines from the
four tissues. For serial passage, xenografts were harvested at the
time of animal sacrifice, cut into bits of approximately 20-25
mm3, and implanted s.c.
Determination of in Vivo Growth. Six to 8-week-old
athymic mice were implanted s.c. with 20-25-mm3 tumor bits
under anesthesia. The three LuCaP xenograft sublines were
implanted in 10 intact male, 10 castrated male, and 10 female
nude mice. Tumor volume was measured weekly as soon as the
tumor was visible using microcalipers. Tumor volume was
estimated using the formula of an ellipsoid (length X width X
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
-10 10 30 50 70
Time (Days)
B
C’)
EE
C)
E
0>
0
E
I-.
C
-10 0 10 20 30 40 50 60 70 80
Time (Days)
1500
C’)EE
C)
E
0
>0
E1-
750
500
250
-10 0 10 20 30 40 50 60 70 80
Fig. I Growth curves for the LuCaP 23 sublines based on tumor
volume (0) and serum PSA levels (s). A, LuCaP 23. 1 ; B, LuCaP 23.8:
and C, LuCaP 23. 12. Time 0 represents first visible tumor growth.
Tumors grow in an exponential fashion with a PSA increase paralleling
tumor growth. PSA production is highest in the LuCaP 23.12 xenograft,which also has the slowest growth rate.
Clinical Cancer Research 1041
4 E. Corey, manuscript in preparation.
A900
I� 800
!. 700
� 600
.2 5000
> 400
� 300
� 200I-
100
height X 0.5236). Tumor volume doubling time was cal-
culated during exponential growth using the formula t[l/21 =
(t, t1)/(LnV2-LnV1).
Determination of Serum PSA and PAP Concentrations.
Blood samples were obtained weekly from the date of implan-
tation by tail vein bleed into a capillary tube. Approximately 40
pi blood were collected using a 0.65-ml microcentrifuge tube
and centrifuged. Fifteen to 20 p.1 serum were recovered and
stored at -20#{176}Cuntil assayed. Serum PSA and PAP concentra-
tions were determined using the automated IMx PSA immuno-
assay system (Abbott Laboratories, Chicago, IL) from I :10
dilutions of 10-pA serum samples in HBSS or diluent supplied
1500 with the IMx kits.
The average PSA index, defined as the increase in serum1250 PSA per unit of tumor volume (ng/ml/mm3), was calculated
1000 � using the equation of the regression plot analysis of tumor
4 volume against serum PSA concentration in the three xe-
750 g� nografts. To determine the half-life of serum PSA in LuCaP
500 E 23.1 mice, PSA values were obtained before and at 2, 6, 12, 24,� 36, 48, and 60 h after excision of LuCaP 23. 1 tumors in seven
250 � mice. A log linear regression model was used to evaluate the
half-life of serum PSA as described previously (18).
0 Effects of Castration. Athymic mice were implanted
with tumor bits s.c. over the right scapular region. Animals with
tumor volume at least 50 mm3 were castrated at 10 weeks
1 500 postimplantation. Castration was performed through a transab-
dominal incision under anesthesia. Tumor volume and serum
I 250 �E PSA were measured weekly after castration. PSA nadir and time
1000 � to PSA nadir were compared to progression status using un--� paired two-tailed t tests on computer software.
750 � Histology. Five-p.m paraffin-embedded tumor sections
E were stained with H&E to determine morphological structure500 � and histological grading according to the method of Gleason
250 � (19). For immunohistochemical studies, 5-�i.m sections weredeparaffinized with Histoclear II (National Diagnostics, Atlanta,
GA) and rehydrated in degressive gradients of ethanol. Standard
avidin/biotin-enhanced immunoperoxidase staining protocols
were used to determine immunoreactivities to pertinent antibod-
ies: the anti-PSA monoclonal antibodies l6-3A2 and 22-8A2
and the antirenal cell carcinoma/antiprostate cancer monoclona!
1 250 ,� antibody A6H, which were generated in our laboratory,4 anti-
:� prostate carcinoma monoclonal antibodies PD-4 1 , TURP-27,1 000 � and 7El i-CS and anti-PAP monoclonal antibody (kindly per-
< formed by Dr. George Wright, Eastern Virginia School of
g� Medicine, Norfolk, VA; Refs. 20-22), and a monoclonal anti-
E body to the androgen receptor (F39.4.l ; Biogenex). Nonspecific
� idiotype-matched monoclonal antibodies were used as negative
C/) controls. Tissue sections were blocked for endogenous peroxi-
dases (1 % H,O2) and biotin prior to the addition of the primary
antibody. In one instance, the anti-PSA monoclonal antibody
16-3A2 was biotinylated, directly avoiding the use of a second
antibody.
Electron Microscopy. Representative 5-I 0-mm3 frag-
ments from throughout the tumor were fixed overnight in a
mixture of 2.5% glutaraldehyde and 2% paraformaldehyde in
cacodylate buffer. Tissue was rinsed in buffer and postfixed in
osmium tetroxide. Dehydration was in graded ethanol solutions
and propylene oxide. Infiltration followed with a mixture of
epoxy (Medcast; Ted Pella, Inc.) and propylene oxide and
embedded in epoxy. One-pm sections were stained with azure
Il/methylene blue. Selected areas were thin sectioned, stained
with urinal acetate and lead citrate, and examined with a Philips
4 10 electron microscope.
5-a Reductase Assay. Northern blot analysis for the
5a-reductase was kindly performed by the laboratory of Dr.
Nicholas Bruchovsky (British Columbia Cancer Institute). Spe-
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
C�)
EE
C)E
0>
0
E
I-.
Fig. 2 An example of differential growth ofthe LuCaP 23 xenograft in
intact male, castrated male, and intact female mice. Data shown are from
the LuCaP 23.1 subline. Growth is slowed in the castrated mice, and no
growth is observed in the female mice.
DAY
1042 LuCaP 23 Prostatic Cancer Xenograft
5 B. J. Williams et al., manuscript in preparation.
Table 2 Tumor take rates in the human prostate cancerxenograft LuCaP 23
LuCaP 23.1 LuCaP 23.8 LuCaP 23.12
l00%(l0/l0)� 90%(9/lO) 67%(6/9)40% (4/10) 50% (5/10) 20% (2/10)
0% (0/10) 0% (0/10) 0% (0/10)
Intact males
Castrated males
Intact females
a Numbers in parentheses, number of successful tumor takes vs.
number implanted.
cific eDNA probes for the Sa-reductase type 1 and 2 isoen-
zymes were hybridized to LuCaP 23 mRNA as described pre-
viously (23). Human BPH mRNA was used as a control.
Chromosomal Analysis. Chromosomal analysis was
performed as a courtesy by Dr. Arthur Brothman (University of
Utah).5 Animals containing xenografts of LuCaP 23.1 passage 7
were sacrificed by cervical dislocation, and the entire tumor was
excised. Tumor cells were mechanically dissociated and washed
three times in HEPES-buffered saline (10 mrvi HEPES, 4 mM
glucose, 3 mM KC1, 130 mrvt NaC1, and 1 m�vt Na2HPO4), and
triplicate specimens were incubated at 37#{176}Cfor 1 2 h, 36 h, and
several days (in attempt to establish an in vitro culture) in RPM!
1640 supplemented with 10% fetal bovine serum (Life Tech-
nologies, Inc.). After the specified times, 0.01 p.g/ml Colcemid
was added for mitotic arrest, and specimens were 0-banded as
described previously (24). FISH using biotinyiated probes for
total human DNA (P5080; Oncor, Gaithersburg, MD), total
mouse DNA (P2930; Oncor), and total human centromere DNA
(P5095; Oncor) was performed on prepared cells using estab-
lished procedures (25).
RESULTS
In Vitro Growth. There was initial outgrowth of epithe-
lioid cells in a confluent monolayer from tumor fragments
harvested at autopsy. PSA was detected in conditioned media of
the tissue culture flasks. Detection of PSA was lost after one to
two passages, and cells became senescent shortly thereafter.
Subsequently, numerous attempts were made to establish per-
manent cell lines from the established xenografts. Cells were
cultured in two different growth medias, in addition to a variety
of plate coatings, supplemented with growth factors and hor-
mones. Cultures were established from the xenografts and be-
haved similar to cultures derived from the initial tumor frag-
ments (e.g. , confluent monolayers of epithelioid cells with
detectable PSA in the conditioned tissue culture media). How-
ever, similar to the primary cultures established from autopsy
specimens, the cells became senescent after one to two passages.
In Matrigel, cells formed spheroid colonies, and cell growth was
maintained for up to 4 weeks, but continuous cultures were not
attainable.
Establishment of the Xenografts. Tumor bits harvested
from two separate retroperitoneal lymph nodes and a liver
metastasis in the donor evolved into three xenograft sublines
(designated LuCaP 23.1, 23.8, and 23.12, respectively), which
are serially transplantable and currently maintained in our ani-
ma! facility. Growth of tumors was first detectable approxi-
mately 55 days postimpiantation and confirmed by the presence
of PSA and PAP in mouse serum. From the time of initial
implantation of tumors in August 1991 until June 1995, the
LuCaP 23.1, 23.8, and 23. 12 xenografts have been passaged 17,
17, and 14 times, respectively. Xenografts are currently pas-
saged approximately every 14 weeks. Once established as a
transplantable xenograft, the growth rates have not changed.
In Vivo Growth. Tumors arising from s.c. implantation
of 20-25-mm3 tumor bits tumors are typically visible by days
30-45. Representative growth curves from the three sublines
are shown in Fig. I . Tumor doubling times for the LuCaP 23. I,
23.8, and 23.12 sublines are approximately 11, 15, and 21 days,
respectively. In the three xenografts, tumor take rates are im-
paired in castrated male mice, and no tumors grow when im-
planted in female mice, as shown in Table 2. Tumor growth is
also decreased in castrated animals (Fig. 2). Growth in castrated
female mice is comparable to that in castrated male mice be-
cause tumor take rates and growth are similar (data not shown).
Determination of Serum PSA and PAP Concentrations.Production of PSA by the LuCaP 23 xenografts proceeds visible
tumor growth by 1-3 weeks and is detectable in the serum
approximately 14-28 days postimplantation. As shown in Fig.
2, serum PSA in intact male mice correlates with tumor volume
in each xenograft (,2 0.6, 0.8, and 0.8 respectively, in LuCaP
23.1, 23.8, and 23.12). The PSA index was 1.27, 1.63, and 5.21
ng/mi/mm3, respectively, in LuCaP 23.1, 23.8, and 23.12 xe-
nografts. Serum PAP concentrations are measurable in mice
bearing the three LuCaP 23 xenograft sublines in a ratio to
serum PSA concentrations of approximately I : 10. After surgical
removal of the xenografts, the average serum PSA half-life was
12.9 ± 0.9 h (r2 0.9). As previously described for the LNCaP
mouse model, PSA clearance in that of LuCaP follows a two-
compartment model of first-order elimination kinetics (18).
Castration-induced Effects. Castration-induced effects
on serum PSA levels and tumor volume in the LuCaP 23. 1 and
23.12 sublines are shown in Figs. 3 and 4. Individual mouse
serum PSA levels and tumor volumes were normalized to the
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
A400
>1Cs
� 300
I� 200Cl)Q.
f� 100
0
B
0
>‘CC
0
0
Fig. 3 A, PSA response to castration (time 0) in
the LuCaP 23. 1 xenograft can be categorized into
three groups: one group exhibits a prolonged re-
sponse to castration (responders; A), a second
group progresses between days 20 and SO (partial
responders; �), and a third group exhibits mini-
mal response to castration (S). B, tumor volume
in the same animals from A. Animals are catego-
rized by PSA response. Note that tumor volume isaffected less by castration than serum PSA.
-10 0 10 20 30 40 50 60 70
Time Post Castration (Days)
- -4- Minimal Response (n=6)
-U -Intermediate Response (n=12)
-fr--Prolonged Response (n=5)
II
400
300
200
100
0-10 10 30 50 70
Time Post Castration (Days)
90
Clinical Cancer Research 1043
C)
E
0>1�0EI-.
values measured at the time of castration. Tumors were assumed
to be hormone refractory when PSA values exceeded those prior
to castration for two subsequent measurements. Using this cri-
teria, xenografts were classified into three groups according to
their length of response to castration. The first group is charac-
terized by a prolonged decrease (>50 days) in serum PSA levels
after castration. The second group exhibits an intermediate
response to castration (PSA decrease enduring 20-50 days
postcastration). The final group of animals demonstrate a mm-
imal decrease (<20 days) in PSA sera levels following castra-
tion. Based on these criteria, in the LuCaP 23. 1 subline a
prolonged response was noted in 5 (22%) of 23, an intermediate
response was noted in 12 (52%) of 23, and a minimal response
was noted in 6 (26%) of 23 animals after castration (Fig. 3). The
LuCaP 23. 12 subline exhibited a greater response to castration
with 6 (50%) of 12 animals demonstrating a prolonged response,
3 (25%) of 12 animals demonstrating an intermediate response,
and 3 (25%) of 12 animals demonstrating a minimal response to
castration (Fig. 4).
Changes in tumor volume are usually smaller than changes
in serum PSA. In general, those animals with the greatest
response to castration by PSA criteria also show the greatest
response by tumor volume criteria. The variability in response to
castration is independent of the source of the xenograft as
tumors derived from the same parent xenograft produced heter-
ogeneous responses to castration.
The decrease in serum PSA correlated with progression in
the animals. As demonstrated in Table 3, those animals with the
longest time to progression also exhibited the greatest decreases
in serum PSA. These decreases in serum PSA were also con-
tinued for longer periods of time in the better-responding tumors
as reflected in the time to reach PSA nadir.
Histology. Histological examination of the metastatic tu-
mors harvested from the patient revealed sheets of cells with
extensive necrosis. No clear gland formation was noted but the
necrosis precluded adequate evaluation. Examination of the
xenografts reveals a glandular pattern which resembled a
Gleason’s grade 3-4 adenocarcinoma, as shown in Fig. 5. The
nuclei of the cells are pleomorphic with large nucleoli. The
luminal cells exhibit a columnar morphology with polarity.
Scattered neuroendocrine cells are visible. Occasional structures
which appear to be apoptotic bodies are also present, which
increase in number following castration. Little stroma is noted
within the xenografts. Presumably, the stroma that is present is
of mouse origin. Extensive necrosis and cystic areas are appar-
ent within the larger xenografts.
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
A400
350
0 300>1
CC0 250
0 200
-. 1504
� 100
50
0.10 0 10 20 30 40 50 60 70 80
Time Post Castration (Days)
90 Fig. 4 A. LuCaP 23.12 xenografts sub-divided into prolonged (>50 days; A),
partial (between 20 and SO days: �), and
minimal (<20 days: O) response to cas-
tration. B, tumor volume responses to cas-
tration in the same animals from A. Ani-
mals are categorized by PSA response.
-10 0 10 20 30 40 50 60 70 80 90
Time Post Castration (Days)
1044 LuCaP 23 Prostatic Cancer Xenograft
B400
� 350
0 300
0 250
C)E 200
.� 150
>� 100
E� 50I-.
0
Table 3 Relationship be tween PSA re sponse and mm or progression
Time to
progression
<So days
Time to
progression
>50 days P
LuCaP 23.1% PSA decreaseTime to PSA nadir
n = 18
55.7 ± 6.5
10.9 ± 1.1
n - S
82.2 ± 5.3
30.8 ± 10.0
=0.05
<0.005
LuCaP23.12
% PSA decreaseTime to PSA nadir
n6
35.5 ± 7.2
10.0 ± 3.6
n=6
91.7 ± 4.655.7 ± 7.7
<0.001<0.001
Results of immunohistochemical studies are summarized in
Table 4. PSA and PAP were consistently detected within the
cytoplasm of all epithelial cells. An example of immunostaining
with an anti-PSA antibody is demonstrated in Fig. 6. Mono-
clonal antibody 7El 1-CS reacted with at least 30% of the cells
of LuCaP 23. 1 but < 10% of the cells within LuCaP 23.8 and
23.12. Monoclonal antibodies PD-4l and TURP-27 did not
show appreciable reactivity in the different xenografts. The
strongest staining against the epithelial cells of the three xe-
nografts was obtained with the monoclonal antibody A6H. Im-
munostaining for the androgen receptor is shown in Fig. 7.
Immunostaining is localized to the tumor cell nuclei with a
heterogeneous staining pattern.
Electron Microscopy. As shown in Fig. 8, LuCaP 23
cells exhibit nuclear polymorphism with prominent nucleoli and
pseudoinclusions of cytoplasm in some nuclei when viewed
with electron microscopy. Desmosomes and plasma membrane
interdigitations are present on the cytoplasmic membrane be-
tween epithelioid cells surrounding the pseudoglandular lumen.
Clusters of round to oval electron-dense structures, which range
in diameter from 0.5 to S p.m. are observed within the cytoplasm
of tumor cells. These structures are ultrastructurally similar to
secondary lysosomes which are abundant in normal and neo-
plastic prostate carcinoma.
5-a Reductase Assay. Northern blot analysis of mRNA
extracted from the xenografts is shown in Fig. 9. The message
for Sa-reductase type I is present in the xenografts, whereas the
message for 5a-reductase type 2 is not detected. Both type 1 and
type 2 transcripts are detectable in control BPH mRNA.
Chromosomal Analysis. Metaphase preparations were
of poor quality, and the mitotic index was low for all specimens.
The most satisfactory mitotic values were obtained following
the 12-h incubation; long-term in vitro cultures could not be
established. A total of 48 metaphases were evaluated following
0-banding with Wright stain. The chromosome number ranged
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 1045
C) Unpublished data.
Fig. 5 A, photomicrograph of a representative portion of a H&E-
stained section of tumor. The tumor is composed of glands that have
lumina of various sizes and shapes and that grow back-to-back with
adjacent tumor glands. B, higher magnification photomicrograph of
tumor demonstrates glandular lumina lined by dysplastic tumor cells
that have nuclei basally oriented away from the lumina.
from 62 to 1 12 chromosomes/cell, with a modal number of 78.
Multiple structural and numerical abnormalities were detected,
and many normal human chromosomes could be identified, but
a definitive karyotype could not be prepared due to poor mor-
phology. FISH analysis of both metaphase and interphase nuclei
using the total human DNA showed diffuse signal, whereas the
total mouse DNA showed no signal (a positive control using
mouse cells showed signal with this probe). Total human cen-
tromere DNA showed strong, distinct signals.
DISCUSSION
In this article, we describe a new prostate cancer xenograft
system, LuCaP 23, which retains two ofthe clinical hallmarks of
prostate carcinoma, androgen responsiveness and PSA produc-
tion. Three different sublines were developed from three differ-
ent metastases harvested from the tumor donor shortly after
death. LuCaP 23. 1 and 23.8 were developed from lymph node
metastases, while LuCaP 23. 12 was developed from a liver
metastasis.
The LuCaP 23 series of xenografts were generated from
metastases in a patient with end stage hormone-resistant prostate
cancer. Xenografts derived from such a patient would be ex-
pected to exhibit androgen-independent behavior. However,
these tumors clearly are androgen sensitive, and many of the
tumors appear to be androgen dependent as demonstrated by a
marked decrease in serum PSA levels and a lesser, but demon-
strable, decrease in tumor size following castration. Investiga-
tors have described variable responses to castration in the most
commonly studied xenograft system, LNCaP. Lim et a!. (26)
reported tumor involution in response to androgen withdrawal in
LNCaP xenografts. On the other hand, Gleave et a!. (27) re-
ported tumor size stabilizes, or may even increase, following
castration. In our experience with LNCaP xenografts, the tumors
do not decrease in size following castration.6 In the PC-82
xenograft, Kyprianou et a!. (28) reported that programmed cell
death and tumor involution occur in response to androgen with-
drawal.
Androgenic effects in the human prostate are primarily
regulated by the interaction of DHT with the androgen receptor.
Androgen receptor is present in most of the LuCaP 23 cell
nuclei as determined using immunostaining. Whether or not
there are mutations of this androgen receptor is not yet known.
DHT is produced from testosterone by the enzyme 5 a-reduc-
tase. The heterogeneity in androgen responsiveness could be
related to heterogeneity in the androgen receptor staining. The
predominant isozyme of 5 a-reductase present in the prostate is
type 2 (29). Interestingly, this isozyme is present in the prostatic
stromal cells, not in the epithelial cells. Prostate epithelial cells
produce the type I isozyme of 5-a-reductase. which is also
present in the skin (30). The presence of the type 1 isozyme in
these xenografts is consistent with the prostatic epithelial origin
of these tumors.
The production of PSA is an important characteristic of
prostate carcinomas and a desirable characteristic of prostate
cancer models. PSA is a Mr 34,000 serine protease produced by
benign and malignant prostate epithelial cells, which is reviewed
elsewhere (3 1 ). This enzyme is responsible for liquefaction of
the seminal coagulum (32). The discovery that serum PSA
levels can be elevated by prostatic neoplasms has made PSA one
of the most valuable tumor markers in all of oncology. PSA has
replaced bone scans and acid phosphatase determinations as the
primary means of monitoring disease status in patients with
prostate cancer. Responses to therapy are commonly described
in terms of changes in PSA levels. The production of PSA by
prostate cancer xenografts is an important characteristic which
confirms prostatic origin of the tumor, suggests the model will
behave in a fashion similar to human prostate cancer, and allows
the modeling of clinical scenarios. Serum PSA levels in mice
bearing the LuCaP 23 xenografts correlate with tumor growth.
Following androgen ablation, the decrease in PSA as well as the
time to reach PSA nadir correlate with time to tumor progres-
sion. Both of these observations model the findings in clinical
prostate cancer. The expression of this marker allows the inves-
tigator to monitor tumor burden when the tumors are not easily
measurable (i.e. , subrenal capsule implantation, orthotopic pros-
tate implantation, extremely small tumors, or metastatic dis-
ease). The PSA production by LuCaP 23 is many fold higher
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
‘/
�:
�. 1..�.
#{149},
-. :. :�#{149};f4-t�:..
: -4�:� -�� �
� . .�,tt
r�/r’f�. �
�j,.a ..�- � .‘ : � � : ,�. �
#{149}:,43� �
- �4r . �
#{149};t�#{149}*
-�
� . I� �‘ �
�.‘t. ,�r_ ,,� �,
� � � z� ,.-..
. 0 *�
� .-.. � 4’?P’#{149}�� . �.1.4 1
�‘, . �1 � � � , I ;� � �;
�; ; � �, � t .�1 �#; j�4���1lI1t� �. �
Fig. 6 A, immunohistochemical staining with anti-PSA monoclonal
antibodies. The cells polarized with intense apical cytoplasmic reactivitywith the anti-PSA antibodies. B, MOPC control.
than that by any of the other prostate cancer xenografts which
are known to produce measurable serum PSA levels.
Because in vitro cultures could not be established, detailed
classical cytogenetic analysis could not be performed. The prep-
arations were acceptable, however, to determine the human
Fig. 7 High magnification of tumor immunostained for androgen re-
ceptor. Immunoreactivity is localized exclusively in the tumor epithelial
cell nuclei.
origin of LuCaP 23.1 by G-banding, and a hypertriploid to
near-tetraploid karyotype was observed. FISH proved to be a
valuable technique for confirmation of these findings, using
total human and total mouse probes. There was no evidence of
any mouse cells in the harvested material evaluated. A more
detailed molecular cytogenetic characterization of these cells is
ongoing and will be presented elsewhere.
The reason for the relative lack of expression of PD4 1,
TURP-27, and 7El 1-CS is unclear. PD-41 is an important
antigen which appears to react with most (65%) human prostatic
carcinomas (20). The TURP-27 antigen is expressed to various
degrees by both benign and malignant prostatic tissue (33). The
7El 1-CS antibody was developed by Horoszewicz et a!. (22) to
an antigen in the prostate cancer cell membrane. The antibody is
used experimentally in radioimmunoscintigraphy and radioim-
munotherapy. One of the most interesting aspects of 7E1 i-CS is
that while it is expressed in most prostate cancer cells, overex-
pression occurs in those tumors exhibiting hormone independ-
ence (34). In the LuCaP 23 series, <30% of the cells exhibit
7E1 1-CS reactivity. It is not known whether these cells represent
a hormonally independent subset of the tumor cells.
Although the characteristics of the three LuCaP 23 sublines
1046 LuCaP 23 Prostatic Cancer Xenograft
Table 4 Results of immunohistochemical staming of the LuCaP 23 sublines
LuCaP 23.1 LuCaP 23.8 LuCa P 23.12
Staininga % Positive” Staining % Positive Staining % Positive
Monoclonal antibodies intensity cells intensity cells intensity cells
Anti-PAP 2-3+ 100 2-3+ 100 2-3+ 100
Anti-PSA16-3A2 3-4+ 100 3-4+ 100 3-4+ 10022-8A2 3-4+ 100 3-4+ 100 3-4+ 100
Antirenal cancer and prostate cancerA6H 4+ 100 4+ 100 4+ 100
Antiprostate carcinoma7E1 1-CS 2-3+ 30 1-2+ SC” 0PD-4l 2-3+ SC 0 2-3+ 1
TURP-27 2-3+ SC 3-4+ SC 2-3 + SCa o, no staining; 1 +, weak staining; 2+, moderate staining; 3+, strong staining; 4+, very strong staining.
b SC scattered cells.
. �.‘s� #{149}
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
.-..( � ...‘ ,
.. � � . // .. � .
“t.’r’� . .
-. � �- -. .‘. � - : �j
. ... .,- .: �...#{231}..,,.. #{149} . . �
� t/ -� � �
.,/� � � �
� 0 A�i1:� � 4t
Fig. 8 Electron micrograph of LuCaP 23.1 xenograft. Many of the
nuclei contain large nucleoli. Numerous electron-dense granules of
uncertain origin are present within the cytoplasm of the cells. X3000.
Type 1 Type 2
2.1 kb- �
2. Okada, K., and Schroeder, F. Human prostatic carcinoma in cell
culture: preliminary report on the development and characterization of
an epithelial cell line. Urol. Res., 2: 1 1 1-121, 1974.
____ A A 1,-k � Stone, K. R., Mickey, D. D., Wunderli, H., Michey, G. H., and-F..,. � Paulson, D. F. Isolation of a human prostate carcinoma cell line (DU-
- - 2.4 kb 145). Intl. J. Cancer, 21: 274-281, 1978.4. Kaighn, M., Shakar Narayan, K., Ohnuki, Y., Lechner, J., and Jones,L. Establishment and characterization of a human prostatic carcinoma
cell line (PC-3). Urology, 17: 16-23, 1979.
LB LB
Clinical Cancer Research 1047
Fig. 9 Northern blot hybridization of LuCaP 23 (L) and human BPH
(B) mRNA with Sa-reductase type 1 (left panel) and type 2 (right
panel)-specific eDNA. The Sa-reductase type I eDNA hybridizes to a
2.1-kb transcript visible in both the LuCaP 23 and the BPH mRNA. The
Su-reductase type 2 eDNA hybridizes to a 2.4- and 4.4-kb transcript
visible only in the BPH lanes.
described in this article are similar, there are apparent differ-
ences in doubling time, PSA production, and hormonal respon-
siveness. LuCaP 23. 12 appears to be the most differentiated of
the tumors by virtue of a slower growth rate, higher PSA
production, and greater response to androgen deprivation. These
differences are to be expected given the heterogeneous nature of
tumors and metastases. This tumor heterogeneity appears to
extend to the individual xenograft since multiple explants of a
single xenograft are noted to have differing responses to castra-
tion. In contrast, xenografts derived from cell lines tend to be
more homogeneous in their behavior because each animal is
inoculated with a cell population representative of the uniform
cell culture population. Although heterogeneous xenografts are
more difficult to study, they nonetheless represent the complex-
ities of clinical prostatic carcinoma, which is a multifocal dis-
ease with different cytogenetic markers expressed by different
tumor foci. This heterogeneity is reflected in the heterogeneous
response to therapy in human prostate carcinoma.
To date, we have been unsuccessful in immortalizing these
cells in tissue culture. The development of such a cell line is
important because it allows study of questions that cannot be
answered in the xenograft model. However, if heterogeneity is
as important as we believe, the transfer from xenograft to cell
line and back to xenograft may result in an inferior model due
to the homogeneity of the new tumor. Similarly, immortaliza-
tion of the cells by viral transfection might change the charac-
teristics of this model. Nonetheless, such efforts are under
consideration.
We believe this new human prostate cancer xenograft
system will provide valuable insight into the biology of prostate
carcinoma. This model has the desirable characteristics of PSA
production, hormonal responsiveness, and a relatively rapid
growth rate. Our hope is that these characteristics wiii allow
study of the changes associated with the progression from an
androgen-dependent to an androgen-independent tumor.
REFERENCES
1. Boring, C. C.. Squires. T. S., and Tong, T. Cancer statistics, 1993.
CA Cancer J. Clin., 43: 7-26, 1993.
S. Horoszewicz, J. S., Leong, S. S., Chu, T. M., Wajsman. Z. L.,
Freedman, M., Papsidero, L., Kim, U., Chai, L. S., Kakati, S., Arya.
S. K.. and Sandberg, A. A. The LnCaP cell line-a new model for
studies on human prostatic carcinoma. Prog. Clin. Biol. Res., 37. 1 15-
132, 1980.
6. Class. F.. and Van Steenbrugge, 0. J. Expression of HLA-like
structures on a permanent human tumor line PC-93. Tissue Antigens.
21: 227-232, 1983.
7. lizumi, T., Yazaki, T.. Kanoh, S., Kondo. I., and Koiso. K. Estab-lishment of a new prostatic carcinoma cell line (TSU-PRI). J. Urol.,
137: 1304-1306, 1987.
8. Muraki, J., Addonizio, J., Choudhury, M., Fischer, J., Eshghi, M.,
Davidian, M., Sharppiro, L., Wilmot, P., Nagamatsu. G.. and Chiao. J.
Establishment of new human prostatic cancer cell (JCA- I ). Invest.
Urol., 36: 79-84, 1990.
9. Gingrich, J., Tucker, J., Walther, P., Day. J.. Poulton, S., and Webb,
K. Establishment and characterization of a new human prostatic carci-noma cell line (DuPro-i). J. Urol.. 146: 915-919. 1991.
10. Narayan. P.. and Dahiya, R. Establishment and characterization of
a human primary prostatic adenocarcinoma line (ND-I). J. Urol.. 148:
1600-1604, 1992.
1 1. Hoehn, W., Schoroeder, F. H., Riemann, J. F., Joebsis, A. C., and
Hermanek, P. Human prostatic adenocarcinoma: some characteristics of
a serially transplantable line in nude mice (PC 82). Prostate. 1: 95-104,1980.
12. Ito, Y. Z., Mashimo, S., Nakazato, Y., and Takikawa, H. Hormone
dependency of a serially transplantable human prostatic cancer
(HONDA) in nude mice. Cancer Res., 45: 5058-5063, 1985.
13. Graham, S. D., Poulton, S. H., Linder, J., Woodard, B. H., Lyles,
K. W., and Paulson, D. F. Establishment ofa long-term adenocarcinomaof the prostate cell line in the nude mouse. Prostate, 7: 369-376, 1985.
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1048 LuCaP 23 Prostatic Cancer Xenograft
14. Hoehn, W., Wagner, M., Riemann, J. F., Hermanek, P., Williams,
E., Walther, R., and Schruffer, R. Prostatic adenocarcinoma PC-EW, anew human tumor line transplantable in nude mice. Prostate, 5: 445-
452, 1984.
15. Van Steenbrugge, 0., Oroen, M., Bolt-de Vries, J., Romijn, J., and
Schroeder, F. Human prostate cancer (PC-82) in nude mice: a model to
study androgen regulated tumor growth. Prog. Clin. Biol. Res., 185A:
23-SO, 1985.
16. Caspo, Z., Brand, K., Walther, R., and Konstantinos, F. Compara-
tive experimental study of the serum prostate specific antigen andprostate acid phosphatase in serially transplantable human prostaticcarcinoma lines in nude mice. J. Urol., 140: 1032-1038, 1988.
17. Harper, M., Sibley, P., Rowlands, A., Buttifant, L., Beacock, C.,
and Griffiths, K. Hormonal modulation of the growth of a new trans-
piantable prostatic cell line in athymic mice. Urol. Res., 14: 1S6A, 1986.
18. Oleave, M. E., Hsieh, J. T., Wu, H., Von Eschenbach, A. C., andChung, L. W. Evaluation of the pharmacokinetics of PSA using theLNCaP tumor model. J. Urol., 147: 694, 1992.
19. Oleason, D. F. Histologic grading in clinical staging of prostatic
carcinoma. In: M. Tannenbaum (ed), Urologic Pathology: The Prostate,pp. 171-179. Philadelphia: Lea & Febiger, 1977.
20. Beckett, M. L., Lipford, 0. B., Haley, C. L., Schellhammer, P. E.,
Wright, 0. L., Jr. Monocional antibody PD41 recognizes an antigenrestricted to prostate adenocarcinomas. Cancer Res., 51: 1326-1333,
1991.
21. Starling, J. J., Sieg, S. M., Beckett, M. L., Wirth, P. R., Wahab, Z.,
Schellhammer, P. F., Ladaga, L. E., Ploeskic, S., and Wright, 0. L., Jr.Human prostate tissue antigens defined by murine monoclonal antibod-ies. Cancer Res., 46: 367-374, 1986.
22. Horoszewicz, J. S., Kawinski, E., and Murphy, 0. P. Monoclonalantibodies to new antigenic marker in epithelial prostatic cells andserum of prostatic cancer patients. Anticancer Res., 7: 927-936, 1987.
23. Kaefer, M., Audia, J., Bruchovsky, N., Goode, R., Hsiao, K.,
Libovitch, I., Kruchinsky, J., Lee, C., Steidle, C., Sutowski, D., andNeubauer, B. Characterization of type I 5-alpha reductase activity inDU145 human prostatic adenocarcinoma cells. J. Steroid Biochem. Mo!.Biol., in press, 1996.
24. Brothman, A. R., Peeh, D. M., Patel, A. M., and McNeai, J. E.
Frequency and pattern of karyotypic abnormalities in human prostate
cancer. Cancer Res., 50: 3795-3803, 1990.
25. Brothman, A., Patel, A., Peehi, D., and Schellhammer, P. Analysisof prostatic tumor cultures using fluorescence in situ hybridization
(FISH). Cancer Genet. Cytogenet., 62: 180-185, 1992.
26. Lim, D. J., Liu, X. L., Sutkowski, D. M., E. J. B., Lee, C.,
Koziowski, J. M. Growth of an androgen-sensitive human prostatecancer cell line, LNCaP, in nude mice. Prostate, 22: 109-118, 1993.
27. Gleave, M. E., Hsieh, J. T., Wu, H. C., von Exchenbach, A. C., andChung, L. W. K. Serum prostate specific antigen levels in mice bearinghuman prostate LNCaP tumors are determined by tumor volume andendocrine and growth factors. Cancer Res., 52: 1598-1605, 1992.
28. Kyprianou, N., English, H., and Isaacs, J. Programmed cell death
during regression of PC-82 human prostate cancer following androgenablation. Cancer Res., 50: 3748-3753, 1990.
29. Thigpen, A. E., Silver, R. I., Ouileyardo, J. M., Casey, M. L.,
McConnell, J. D., and Russell, D. W. Tissue distribution and ontogenyof steroid S aipha-reductase isoenzyme expression. J. Clin. Invest., 92:
903-910, 1993.
30. Harris, 0., Azzolina, B., Baginsky, W., Cimis, 0., Rasmusson,
0. H., Toiman, R. L., Raetz, C. R. H., and Ellsworth, K. Identificationand selective inhibition of an isozyme of steroid S aipha-reductase in
human scalp. Proc. Nat!. Acad. Sci. USA, 89: 10787-10791, 1992.
31. Ellis, W. J., and Brawer, M. K. The role of tumor markers in the
diagnosis and treatment of prostate cancer. In: H. Lepor and R. K.Lawson (eds.), Prostate Diseases, pp. 276-291. Philadelphia: W. B.
Saunders Company, 1993.
32. Liija, H. A kaliikrein-like serine protease in prostatic fluid cleaves
the predominant seminal vesicie protein. J. Clin. Invest., 76: 1899-1903, 1985.
33. Wright, 0. L., Jr., Beckett, M. L., Lipford, 0. B., Haley, C. L., and
Schelihammer, P. F. A novel prostate carcinoma-associated glycopro-tein complex (PAC) recognized by monoclonal antibody Turp-27. mt. J.
Cancer, 47: 717-725, 1991.
34. Israeli, R. S., Powell, C. T., Corr, J. 0., Fair, W. R., and Heston,
W. D. W. Expression of the prostate-specific membrane antigen. Cancer
Res., 54: 1807-1811, 1994.
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1996;2:1039-1048. Clin Cancer Res W J Ellis, R L Vessella, K R Buhler, et al. xenograft: LuCaP 23.prostate-specific antigen-producing prostatic carcinoma Characterization of a novel androgen-sensitive,
Updated version
http://clincancerres.aacrjournals.org/content/2/6/1039
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://clincancerres.aacrjournals.org/content/2/6/1039To request permission to re-use all or part of this article, use this link
Research. on November 16, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from